A pretreatment multiparametric MRI-based radiomics-clinical machine learning model for predicting radiation-induced temporal lobe injury in patients with nasopharyngeal carcinoma.

BACKGROUND: To establish and validate a machine learning model using pretreatment multiparametric magnetic resonance imaging-based radiomics data with clinical data to predict radiation-induced temporal lobe injury (RTLI) in patients with nasopharyngeal carcinoma (NPC) after intensity-modulated radiotherapy (IMRT). METHODS: Data from 230 patients with NPC who received IMRT (130 with RTLI and 130 without) were randomly divided into the training (n = 161) and validation cohort (n = 69) with a ratio of 7:3. Radiomics features were extracted from pretreatment apparent diffusion coefficient (ADC) map, T2-weighted imaging (T2WI), and CE-T1-weighted imaging (CE-T1WI). T-test, spearman rank correlation, and least absolute shrinkage and selection operator (LASSO) algorithm were employed to identify significant radiomics features. Clinical features were selected with univariate and multivariate analyses. Radiomics and clinical models were constructed using multiple machine learning classifiers, and a clinical-radiomics nomogram that combined clinical with radiomics features was developed. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were drawn to compare and verify the predictive performances of the clinical model, radiomics model, and clinical-radiomics nomogram. RESULTS: A total of 5064 radiomics features were extracted, from which 52 radiomics features were selected to construct the radiomics signature. The AUC of the radiomics signature based on multiparametric MRI was 0.980 in the training cohort and 0.969 in the validation cohort, outperforming the radiomics signature only based on T2WI and CE-T1WI (p < 0.05), which highlighted the significance of the DWI sequence in the prediction of temporal lobe injury. The area under the curve (AUC) of the clinical model was 0.895 in the training cohort and 0.905 in the validation cohort. The nomogram, which integrated radiomics and clinical features, demonstrated an impressive AUC value of 0.984 in the validation set; however, no statistically significant difference was observed compared to the radiomics model. The calibration curve and decision curve analysis of the nomogram demonstrated excellent predictive performance and clinical feasibility. CONCLUSIONS: The clinical-radiomics nomogram, integrating clinical features with radiomics features derived from pretreatment multiparametric MRI, exhibits compelling predictive performance for RTLI in patients diagnosed with NPC.

Metabolic phenotyping reveals an emerging role of ammonia abnormality in Alzheimer's disease.

The metabolic implications in Alzheimer's disease (AD) remain poorly understood. Here, we conducted a metabolomics study on a moderately aging Chinese Han cohort (n = 1397; mean age 66 years). Conjugated bile acids, branch-chain amino acids (BCAAs), and glutamate-related features exhibited strong correlations with cognitive impairment, clinical stage, and brain amyloid-ß deposition (n = 421). These features demonstrated synergistic performances across clinical stages and subpopulations and enhanced the differentiation of AD stages beyond demographics and Apolipoprotein E ε4 allele (APOE-ε4). We validated their performances in eight data sets (total n = 7685) obtained from Alzheimer's Disease Neuroimaging Initiative (ADNI) and Religious Orders Study and Memory and Aging Project (ROSMAP). Importantly, identified features are linked to blood ammonia homeostasis. We further confirmed the elevated ammonia level through AD development (n = 1060). Our findings highlight AD as a metabolic disease and emphasize the metabolite-mediated ammonia disturbance in AD and its potential as a signature and therapeutic target for AD.

Long-term follow-up of protective effects on salivary and swallowing structures and improvement of late xerostomia and dysphagia by level IIb optimisation in clinical target volume of nasopharyngeal carcinoma.

BACKGROUND: This study aimed to assess the long-term effect of level IIb clinical target volume (CTV) optimisation on survival, xerostomia, and dysphagia in patients with nasopharyngeal carcinoma (NPC). METHODS: Clinical data of 415 patients with NPC treated with intensity-modulated radiotherapy between December 2014 and October 2018 were retrospectively analysed. The patients were categorised into modified and comparison groups. Late xerostomia and dysphagia were evaluated using Radiation Therapy Oncology Group/European Organisation for Research and Treatment of Cancer scoring. Survival analysis was performed using the Kaplan-Meier method. Differences in late toxicity and dose parameters between both groups were compared. Prognostic factors for survival and late toxicity were assessed using regression analyses. RESULTS: Patients in the modified group developed late xerostomia and dysphagia less frequently than those in the comparison group did (P < 0.001). The mean dose (Dmean) and V26 of parotid glands; Dmean and V39 of submandibular glands; and Dmean of sublingual glands, oral cavity, larynx, and superior, middle, and lower pharyngeal constrictor muscles were lower in the modified group than those in the comparison group (all P < 0.001). Both groups had no significant differences in overall, local recurrence-free, distant metastasis-free, or progression-free survival. The Dmean of the parotid and sublingual glands was a risk factor for xerostomia. The Dmean of the parotid and sublingual glands and middle pharyngeal constrictor muscle was a risk factor for dysphagia. CONCLUSIONS: Level IIb optimisation in NPC patients who meet certain criteria specially the exclusion of positive retropharyngeal nodes treated with intensity-modulated radiotherapy has the potential to better protect the salivary and swallowing structures, decreasing the development of late radiation-induced xerostomia and dysphagia while maintaining long-term survival.

Blockage of mechanosensitive Piezo1 channel alleviates the severity of experimental malaria-associated acute lung injury.

BACKGROUND: Malaria-associated acute lung injury (MA-ALI) is a well-recognized clinical complication of severe, complicated malaria that is partly driven by sequestrations of infected red blood cells (iRBCs) on lung postcapillary induced impaired blood flow. In earlier studies the mechanosensitive Piezo1 channel emerged as a regulator of mechanical stimuli, but the function and underlying mechanism of Piezo1 impacting MA-ALI severity via sensing the impaired pulmonary blood flow are still not fully elucidated. Thus, the present study aimed to explore the role of Piezo1 in the severity of murine MA-ALI. METHODS: Here, we utilized a widely accepted murine model of MA-ALI using C57BL/6 mice with Plasmodium berghei ANKA infection and then added a Piezo1 inhibitor (GsMTx4) to the model. The iRBC-stimulated Raw264.7 macrophages in vitro were also targeted with GsMTx4 to further explore the potential mechanism. RESULTS: Our data showed an elevation in the expression of Piezo1 and number of Piezo1+-CD68+ macrophages in lung tissues of the experimental MA-ALI mice. Compared to the infected control mice, the blockage of Piezo1 with GsMTx4 dramatically improved the survival rate but decreased body weight loss, peripheral blood parasitemia/lung parasite burden, experimental cerebral malaria incidence, total protein concentrations in bronchoalveolar lavage fluid, lung wet/dry weight ratio, vascular leakage, pathological damage, apoptosis and number of CD68+ and CD86+ macrophages in lung tissues. This was accompanied by a dramatic increase in the number of CD206+ macrophages (M2-like subtype), upregulation of anti-inflammatory cytokines (e.g. IL-4 and IL-10) and downregulation of pro-inflammatory cytokines (e.g. TNF-α and IL-1ß). In addition, GsMTx4 treatment remarkably decreased pulmonary intracellular iron accumulation, protein level of 4-HNE (an activator of ferroptosis) and the number of CD68+-Piezo1+ and CD68+-4-HNE+ macrophages but significantly increased protein levels of GPX4 (an inhibitor of ferroptosis) in experimental MA-ALI mice. Similarly, in vitro study showed that the administration of GsMTx4 led to a remarkable elevation in the mRNA levels of CD206, IL-4, IL-10 and GPX-4 but to a substantial decline in CD86, TNF-α, IL-1ß and 4-HNE in the iRBC-stimulated Raw264.7 cells. CONCLUSIONS: Our findings indicated that blockage of Piezo1 with GsMTx4 alleviated the severity of experimental MA-ALI in mice partly by triggering pulmonary macrophage M2 polarization and subsequent anti-inflammatory responses but inhibited apoptosis and ferroptosis in lung tissue. Our data suggested that targeting Piezo1 in macrophages could be a promising therapeutic strategy for treating MA-ALI.

Serum Bile Acids Improve Prediction of Alzheimer's Progression in a Sex-Dependent Manner.

Sex disparities in serum bile acid (BA) levels and Alzheimer's disease (AD) prevalence have been established. However, the precise link between changes in serum BAs and AD development remains elusive. Here, authors quantitatively determined 33 serum BAs and 58 BA features in 4 219 samples collected from 1 180 participants from the Alzheimer's Disease Neuroimaging Initiative. The findings revealed that these BA features exhibited significant correlations with clinical stages, encompassing cognitively normal (CN), early and late mild cognitive impairment, and AD, as well as cognitive performance. Importantly, these associations are more pronounced in men than women. Among participants with progressive disease stages (n = 660), BAs underwent early changes in men, occurring before AD. By incorporating BA features into diagnostic and predictive models, positive enhancements are achieved for all models. The area under the receiver operating characteristic curve improved from 0.78 to 0.91 for men and from 0.76 to 0.83 for women for the differentiation of CN and AD. Additionally, the key findings are validated in a subset of participants (n = 578) with cerebrospinal fluid amyloid-beta and tau levels. These findings underscore the role of BAs in AD progression, offering potential improvements in the accuracy of AD prediction.

The Impact of Drain Placement on Postoperative Complications in Bariatric Surgery: A Systematic Review and Meta-Analysis.

Obesity in individuals can have consequences ranging from metabolically healthy obesity to serious morbidities and reduce the quality and duration of life. A meta-analysis was conducted to assess the role of abdominal drainage on postoperative complications after bariatric surgery. PubMed, Embase, and the Cochrane Library were systematically searched for eligible studies. The results revealed that abdominal drainage was associated with surgical complications, with a pooled odds ratio (OR) of 1.70 (P < .001), but not associated with wound infection (OR: 1.04; P = .762). Associations with surgical complications were mainly detected from retrospective cohort studies. The use of abdominal drainage showed associations with death (OR: 1.68; P < .001) and reoperation (OR: 1.49; P < .001). These findings revealed that abdominal drainage during bariatric surgery was associated with surgical complications, death, and reoperation. These results should be taken with caution since randomized controlled trials and retrospective studies were analyzed together.

Integration of pre-trained protein language models into geometric deep learning networks.

Geometric deep learning has recently achieved great success in non-Euclidean domains, and learning on 3D structures of large biomolecules is emerging as a distinct research area. However, its efficacy is largely constrained due to the limited quantity of structural data. Meanwhile, protein language models trained on substantial 1D sequences have shown burgeoning capabilities with scale in a broad range of applications. Several preceding studies consider combining these different protein modalities to promote the representation power of geometric neural networks but fail to present a comprehensive understanding of their benefits. In this work, we integrate the knowledge learned by well-trained protein language models into several state-of-the-art geometric networks and evaluate a variety of protein representation learning benchmarks, including protein-protein interface prediction, model quality assessment, protein-protein rigid-body docking, and binding affinity prediction. Our findings show an overall improvement of 20% over baselines. Strong evidence indicates that the incorporation of protein language models' knowledge enhances geometric networks' capacity by a significant margin and can be generalized to complex tasks.

Pseudoprogression in the era of immunotherapy-based strategies for recurrent head and neck squamous cell carcinoma achieving complete response: A case report.

RATIONALE: In the last few years, treatment of head and neck squamous cell carcinoma (HNSCC) has been enhanced by the emergence of immunotherapy. A biological phenomenon unique to immunotherapy is pseudoprogression, an increase in tumor burden or the appearance of a new lesion subsequently followed by tumor regression. PATIENT CONCERNS: A 78-year-old man complaining of a lump (6*4 cm) gradually swelling on the right side of his neck with recurrent buccal mucosa squamous cell carcinoma presented to our institution. Two months prior, he received resection of the buccal lesion but refused suggested adjuvant chemoradiotherapy after the operation. DIAGNOSES: Recurrent buccal mucosa squamous cell carcinoma. INTERVENTIONS: Induction immunotherapy was initiated, followed by a new node appearing on the surface of the neck mass. We considered the presence of pseudoprogression and continued with immunotherapy. The patient received immunotherapy combined with chemotherapy and intensity-modulated radiation therapy (IMRT) consecutively. OUTCOMES: The patient experienced an excellent recovery with the disappearance of pain and the lump, along with return of a healthy appetite, weight gain and positive outlook. Complete response (CR) was also noted by magnetic resonance imaging (MRI) scan, with the upper right neck mass significantly retreated to unclear display. The patient is still alive with stable, asymptomatic disease at the time of this writing. LESSONS: These results provide confidence in the safety and efficacy of radical chemo-radio-immunotherapy for the treatment of recurrent, unresectable or metastatic HNSCC.

The Warburg effect modulates DHODH role in ferroptosis: a review.

Ferroptosis is an iron-dependent regulated cell death that suppresses tumor growth. It is activated by extensive peroxidation of membrane phospholipids caused by oxidative stress. GPX4, an antioxidant enzyme, reduces these peroxidized membrane phospholipids thereby inhibiting ferroptosis. This enzyme has two distinct subcellular localization; the cytosol and mitochondria. Dihydroorotate dehydrogenase (DHODH) complements mitochondrial GPX4 in reducing peroxidized membrane phospholipids. It is the rate-limiting enzyme in de novo pyrimidine nucleotide biosynthesis. Its role in ferroptosis inhibition suggests that DHODH inhibitors could have two complementary mechanisms of action against tumors; inhibiting de novo pyrimidine nucleotide biosynthesis and enhancing ferroptosis. However, the link between mitochondrial function and ferroptosis, and the involvement of DHODH in the ETC suggests that its role in ferroptosis could be modulated by the Warburg effect. Therefore, we reviewed relevant literature to get an insight into the possible effect of this metabolic reprogramming on the role of DHODH in ferroptosis. Furthermore, an emerging link between DHODH and cellular GSH pool has also been highlighted. These insights could contribute to the rational design of ferroptosis-based anticancer drugs. Video Abstract.

The development of a self-management evaluation scale for elderly adults with hypertension based on the capability, opportunity, and motivation-behaviour (COM-B) model.

BACKGROUND: Using accurate assessment tools to assess patients in clinical practice is important to mining influencing factors and implementing interventions. However, most evaluation tools for the self-management of elderly patients with hypertension lack a theoretical basis and wide applicability, which makes the intervention effect insignificant. METHODS: Based on the Capability, Opportunity, and Motivation-Behaviour (COM-B) model, combined with literature review and qualitative research, a questionnaire item pool was initially formulated; then the initial items were screened and adjusted through expert consultation and pre-testing to form an initial scale. A field survey of 450 elderly hypertensive patients was then performed using the initial scale to test the reliability and validity of the scale. Cronbach's alpha, test-retest reliability and composite reliability were used to test the reliability of the scale, and the validity of the scale was evaluated from two aspects: content validity and construct validity. The evaluation results of the content validity of the scale by experts were used as the content validity index; the results of exploratory factor analysis and confirmatory factor analysis were used as the structural validity index to further verify the model structure of the scale and develop a formal scale. RESULTS: The final self-management scale included 4 dimensions and 33 items. The Scale-Content Validity Index was 0.920. Exploratory factor analysis extracted four factors that explained 71.3% of the total variance. Cronbach's alpha of the formal scale was 0.867, test-retest reliability was 0.894, and composite reliability of the 4 dimensions were within 0.943 ~ 0.973. Confirmatory factor analysis showed the scale had good construct validity. CONCLUSIONS: The Self-management Capability, Support and Motivation-Behaviour scale for elderly hypertensive patients has good reliability and validity, providing a tool for medical staff to evaluate the self-management level of elderly hypertensive patients.

Homophily-Enhanced Self-Supervision for Graph Structure Learning: Insights and Directions.

Graph neural networks (GNNs) have recently achieved remarkable success on a variety of graph-related tasks, while such success relies heavily on a given graph structure that may not always be available in real-world applications. To address this problem, graph structure learning (GSL) is emerging as a promising research topic where task-specific graph structure and GNN parameters are jointly learned in an end-to-end unified framework. Despite their great progress, existing approaches mostly focus on the design of similarity metrics or graph construction, but directly default to adopting downstream objectives as supervision, which lacks deep insight into the power of supervision signals. More importantly, these approaches struggle to explain how GSL helps GNNs, and when and why this help fails. In this article, we conduct a systematic experimental evaluation to reveal that GSL and GNNs enjoy consistent optimization goals in terms of improving the graph homophily. Furthermore, we demonstrate theoretically and experimentally that task-specific downstream supervision may be insufficient to support the learning of both graph structure and GNN parameters, especially when the labeled data are extremely limited. Therefore, as a complement to downstream supervision, we propose homophily-enhanced self-supervision for GSL (HES-GSL), a method that provides more supervision for learning an underlying graph structure. A comprehensive experimental study demonstrates that HES-GSL scales well to various datasets and outperforms other leading methods. Our code will be available in https://github.com/LirongWu/Homophily-Enhanced-Self-supervision.

Beyond Homophily and Homogeneity Assumption: Relation-Based Frequency Adaptive Graph Neural Networks.

Graph neural networks (GNNs) have been playing important roles in various graph-related tasks. However, most existing GNNs are based on the assumption of homophily, so they cannot be directly generalized to heterophily settings where connected nodes may have different features and class labels. Moreover, real-world graphs often arise from highly entangled latent factors, but the existing GNNs tend to ignore this and simply denote the heterogeneous relations between nodes as binary-valued homogeneous edges. In this article, we propose a novel relation-based frequency adaptive GNN (RFA-GNN) to handle both heterophily and heterogeneity in a unified framework. RFA-GNN first decomposes an input graph into multiple relation graphs, each representing a latent relation. More importantly, we provide detailed theoretical analysis from the perspective of spectral signal processing. Based on this, we propose a relation-based frequency adaptive mechanism that adaptively picks up signals of different frequencies in each corresponding relation space in the message-passing process. Extensive experiments on synthetic and real-world datasets show qualitatively and quantitatively that RFA-GNN yields truly encouraging results for both the heterophily and heterogeneity settings. Codes are publicly available at: https://github.com/LirongWu/RFA-GNN.

Phototrophic Biofilms Transform Soil-Dissolved Organic Matter Similarly Despite Compositional and Environmental Differences.

Dissolved organic matter (DOM) is the most reactive pool of organic carbon in soil and one of the most important components of the global carbon cycle. Phototrophic biofilms growing at the soil-water interface in periodically flooding-drying soils like paddy fields consume and produce DOM during their growth and decomposition. However, the effects of phototrophic biofilms on DOM remain poorly understood in these settings. Here, we found that phototrophic biofilms transformed DOM similarly despite differences in soil types and initial DOM compositions, with stronger effects on DOM molecular composition than soil organic carbon and nutrient contents. Specifically, growth of phototrophic biofilms, especially those genera belonging to Proteobacteria and Cyanobacteria, increased the abundance of labile DOM compounds and richness of molecular formulae, while biofilm decomposition decreased the relative abundance of labile components. After a growth and decomposition cycle, phototrophic biofilms universally drove the accumulation of persistent DOM compounds in soil. Our results revealed how phototrophic biofilms shape the richness and changes in soil DOM at the molecular level and provide a reference for using phototrophic biofilms to increase DOM bioactivity and soil fertility in agricultural settings.

Depletion of SOD2 enhances nasopharyngeal carcinoma cell radiosensitivity via ferroptosis induction modulated by DHODH inhibition.

BACKGROUND: Recurrence due to the development of radioresistance remains a major challenge in the clinical management of nasopharyngeal carcinoma. The objective of this study was to increase the sensitivity of nasopharyngeal carcinoma cells to ionizing radiation by enhancing oxidative stress and ferroptosis caused by disrupting the mitochondrial anti-oxidant enzyme system. METHODS: Oxidative stress cell model was constructed by SOD2 knockdown using shRNA. The expression and activity of DHODH was suppressed by siRNA and brequinar in SOD2 depleted cells. Protein levels were determined by western blotting and ferroptosis was assessed by C11 BODIPY and malondialdehyde assay. Cell viability was evaluated using CCK-8 assay while radiotoxicity was assessed by colony formation assay. Cellular ATP level was determined by ATP assay kits, ROS was determined by DCFD and DHE, while mitochondrial oxygen consumption was determined by seahorse assay. Data were analyzed by two-tailed independent t-test. RESULTS: Radiation upregulated SOD2 expression and SOD2 depletion increased cellular O2.-, malondialdehyde, and the fluorescence intensity of oxidized C11 BODIPY. It also resulted in mitochondrial damage. Its depletion decreased colony formation both under ionizing and non-ionizing radiation conditions. The ferroptosis inhibitor, deferoxamine, rescued cell viability and colony formation in SOD2 depleted cells. Cellular level of malondialdehyde, fluorescence intensity of oxidized C11 BODIPY, O2.- level, ATP, and mitochondrial oxygen consumption decreased following DHODH inhibition in SOD2 depleted cells. Cell viability and colony formation was rescued by DHODH inhibition in SOD2 depleted cells. CONCLUSION: Inducing oxidative stress by SOD2 inhibition sensitized nasopharyngeal carcinoma cells to ionizing radiation via ferroptosis induction. This was found to be dependent on DHODH activity. This suggests that DHODH inhibitors should be used with caution during radiotherapy in nasopharyngeal carcinoma patients.

Combined systematic pharmacology and urine metabonomics to study the therapeutic mechanism of type 2 diabetic treated with the herbal pair of Salvia miltiorrhiza Bunge and Pueraria montana var. lobata (Willd.) Sanjappa & Pradeep.

BACKGROUND: The herbal pair of Salvia miltiorrhiza Bunge and Pueraria montana var. lobata (Willd.) Sanjappa & Pradeep (DG) is commonly used in the treatment of type 2 diabetes (T2DM) in traditional Chinese medicine (TCM). The drug pair DG was designed by Dr. Zhu chenyu to improve the treatment of T2DM. AIM: This study combined with systematic pharmacology and urine metabonomics to explore the mechanism of DG in the treatment of T2DM. METHODS: The therapeutic effect of DG on T2DM was evaluated by fasting blood glucose (FBG) and biochemical indexes. Systematic pharmacology was used to screen the active components and targets that may be related to DG. Metabonomics was established to find urinary metabolites and pathways that may be induced by DG. Finally, integrate the results of these two parts for mutual verification. RESULTS: FBG and biochemical indexes showed that DG could reduce FBG and adjust the related biochemical indexes. Metabolomics analysis indicated that 39 metabolites were related to DG for T2DM treatment. In addition, systematic pharmacology showed compounds and potential targets which were associated with DG. Finally, 12 promising targets were selected as targets for T2DM therapy by integrating the results. CONCLUSION: The combination of metabonomics and systematic pharmacology based on LC-MS is feasible and effective, which provides strong support for exploring the effective components and pharmacological mechanism of TCM.

N6-methyladenosine modification in 18S rRNA promotes tumorigenesis and chemoresistance via HSF4b/HSP90B1/mutant p53 axis.

Aberrant N6-methyladenosine (m6A) modification on mRNA is correlated with cancer progression. However, the role of m6A on ribosomal RNA (rRNA) in cancer remains poorly understood. Our current study reveals that METTL5/TRMT112 and their mediated m6A modification at the 18S rRNA 1832 site (m6A1832) are elevated in nasopharyngeal carcinoma (NPC) and promote oncogenic transformation in vitro and in vivo. Moreover, loss of catalytic activity of METTL5 abolishes its oncogenic functions. Mechanistically, m6A1832 18S rRNA modification facilitates the assembly of 80S ribosome via bridging the RPL24-18S rRNA interaction, therefore promoting the translation of mRNAs with 5' terminal oligopyrimidine (5' TOP) motifs. Further mechanistic analysis reveals that METTL5 enhances HSF4b translation to activate the transcription of HSP90B1, which binds with oncogenic mutant p53 (mutp53) protein and prevents it from undergoing ubiquitination-dependent degradation, therefore facilitating NPC tumorigenesis and chemoresistance. Overall, our findings uncover an innovative mechanism underlying rRNA epigenetic modification in regulating mRNA translation and the mutp53 pathway in cancer.

Efficacy and toxicity of Iodine-125 seed implantation for lymph node recurrence secondary to esophageal cancer after radiotherapy: a multicenter retrospective study.

BACKGROUND/OBJECTIVE: This multicenter study aimed to explore the efficacy and toxicity of radioactive Iodine-125 seed implantation for lymph node recurrence in patients with esophageal cancer after external radiotherapy. METHODS: Clinical data of eligible patients from 5 centers in China were retrospectively reviewed. A total of 126 patients between January 2016 and March 2019 were included. The median interval between previous radiotherapy and radioactive Iodine-125 seed implantation was calculated. The target volume was 2.1-128.1 cm3 (median, 22.2 cm3) and the median postoperative D90 is 120.6 Gy (range, 101.7-192). Short-term efficacy of tumor response, the long-term efficacy of local progression-free survival (LRFS) and overall survival (OS), and treatment-related toxicity were reported. RESULTS: For tumor response, 37 (29.4%), 51 (40.5%), 14 (11.1%), and 24 (19.0%) patients achieved complete response, partial response, stable disease and progressive disease, respectively. The 1-, 2- and 3-year LPFS and OS rates were 48.8%, 23.0% and 15.9%, and 80.2%, 38.8%, and 24.5%, respectively. Multivariate analysis identified Karnofsky performance status (P = 0.041) and tumor response (P = 0.049) as independent prognostic factors for LPFS; initial tumor stage (P = 0.034), lesion volume (P = 0.017), and tumor response (P = 0.004) as independent prognostic factors for OS. In total, 77 (61.1%) patients suffered from skin reactions and the incidence of grade 3-5 skin toxicity was 5.6% (7/126). CONCLUSION: Radioactive Iodine-125 seed implantation seems efficient with acceptable toxicity for the treatment of lymph node recurrence secondary to esophageal cancer. A head-to-head study is needed to further evaluate the survival benefit.

Prognosis of laparoscopic surgery for colorectal cancer in middle-aged patients.

Background: The prognosis of middle-aged patients with colorectal cancer (CRC) treated by laparoscopic resection (LR) is unclear. This study aimed to evaluate the survival outcomes of LR compared with open resection (OR) for middle-aged patients with CRC. Patients and Methods: This retrospective cohort study used the data from a database of all consecutive colorectal resections performed between January 2009 and December 2017. Propensity score matching (PSM) was performed to handle the selection bias based on age, gender, body mass index, tumour location, AJCC stage and admission year. Univariate and multivariate COX regression model was used to identify risk factors of overall survival (OS) and disease-free survival (DFS). Results: After PSM, 154 patients were included in each group. Compared with the OR group in the total cohort, there were better survival outcomes in the LR group for 5-year OS and 5-year DFS (both P < 0.001). These differences were observed for Stage II and III diseases and for all CRC, irrespective of location. The multivariate analysis showed that tumour ≥5 cm (hazard ratio [HR] = 1.750, 95% confidence interval [CI]: 1.026-2.986, P = 0.040), Stage III (HR = 14.092, 95% CI: 1.894-104.848, P = 0.010) and LR (HR = 0.300, 95% CI: 0.160-0.560, P < 0.001) were independently associated with OS. Pre-operative carcinoembryonic antigen ≥5 ng/ml (HR = 3.954, 95% CI: 1.363-11.473, P = 0.011), Stage III (HR = 6.206, 95% CI: 1.470-26.200, P = 0.013) and LR (HR = 0.341, 95% CI: 0.178-0.653, P = 0.001) were independently associated with DFS. Conclusions: In middle-aged patients with CRC, LR achieves better survival than OR. Complications are similar, except for less blood loss and shorter post-surgical hospital stay with LR.

Deep Clustering and Visualization for End-to-End High-Dimensional Data Analysis.

High-dimensional data analysis for exploration and discovery includes two fundamental tasks: deep clustering and data visualization. When these two associated tasks are done separately, as is often the case thus far, disagreements can occur among the tasks in terms of geometry preservation. Namely, the clustering process is often accompanied by the corruption of the geometric structure, whereas visualization aims to preserve the data geometry for better interpretation. Therefore, how to achieve deep clustering and data visualization in an end-to-end unified framework is an important but challenging problem. In this article, we propose a novel neural network-based method, called deep clustering and visualization (DCV), to accomplish the two associated tasks end-to-end to resolve their disagreements. The DCV framework consists of two nonlinear dimensionality reduction (NLDR) transformations: 1) one from the input data space to latent feature space for clustering and 2) the other from the latent feature space to the final 2-D space for visualization. Importantly, the first NLDR transformation is mainly optimized by one Clustering Loss, allowing arbitrary corruption of the geometric structure for better clustering, while the second NLDR transformation is optimized by one Geometry-Preserving Loss to recover the corrupted geometry for better visualization. Extensive comparative results show that the DCV framework outperforms other leading clustering-visualization algorithms in terms of both quantitative evaluation metrics and qualitative visualization.